In the transfer of packet data, the data packet usually contains not only the user data to be transferred but also overhead data, such as the identification (address) of the receiving party, on the basis of which the transfer network is able to route the data packet to its destination. However, standard addresses used in data communication systems (telephone numbers, subscriber identities, equipment identities, etc.) are relatively long so that they require packet transfer capacity which otherwise could be utilized for the transfer of user data. This problem is further aggravated when very short data packets have to be used, e.g., due to the characteristics of the transmission channel. In such very short data packets, standard addresses would use almost all of the packet transfer capacity. In the worst case, the transfer capacity of the data packet is not sufficient to accommodate the address. However, as mentioned above, each data packet in a normal packet data transfer should contain address data in order that the system can identify and route the packets, especially in cases where there are several subscribers transmitting packet data simultaneously on the same channel.